big data analytics in transportation systems management ... · 3. explain the value of analytics in...
TRANSCRIPT
Big Data Analytics in Transportation Systems
Management and OperationsBob McQueen, Bob McQueen and Associates, Orlando, Florida
Petros Xanthopoulos, Stetson University, DeLand, Florida
v3
Topics
• Instructional objectives• Transportation Systems Management and Operations (TSM&O)• TSM&O from an analytics perspective• Predictive and prescriptive analytics• Machine learning and big data• Summary• Discussion
Instructional objectives• At the end of this webinar, you should be able to:
1. Provide an overview of TSM&O2. Define the challenges associated with TSM&O3. Explain the value of analytics in TSM&O4. Define Big Data5. Define big data analytics and relevance to transportation6. Explain the value of Use Cases7. Define an effective approach to Smart Data Management8. Explain predictive and prescriptive analytics9. Describe the relevance of machine learning to TSM&O
Transportation Systems Management and Operations
(TSM&O)Bob McQueen
Transportation Management and Operations• Traffic incident management• Traffic signal coordination• Freeway management • Management of multi-modal transportation systems• Full spectrum from planning to maintenance: transportation as a single system
Project managementProject deliveryTestingCommissioningPartnership management
Define projectsSelect technologyEstimate costDevelop design conceptsDevelop detailed design
Monitor statusCollect dataDevelop informationBuild intelligenceDefine strategiesImplement strategies
Develop maintenance policiesMonitor device statusIdentify intervention pointsAssess device performance
Design Build Operate Maintain
Match supply and demandExplore alternativesUnderstand effectsDevelop results-driven investment programs
Plan
Transportation as a Single System• What is a system?
• It has clarity of purpose• It is connected together• We can find out its status at any
given time• It can adapt to changes in the
environment• “Single system” also includes
alignment between planning, design, project delivery, operations, and maintenance
6
Paraphrased from the speech by Samuel J. Palmisano, Intelligent Transportation Society of America, 2010 Annual Meeting & Conference, Houston, Texas, May 5, 2010
TSM&O Analytics Examples
7
Services AnalyticsAsset and maintenance management Asset performance index, asset maintenance standards compliance measure, optimal intervention point analytic
Connected vehicleLane changes per mile, steering angle compared to road geometry, brake applications per mile, driving turbulence index, minutes per trip, trip time reliability index, no of stops per trip
Connected, involved citizens Citizens awareness levels index, citizens satisfaction levels
Integrated electronic paymentTransit revenue per passenger, transit seat utilization, toll revenue per vehicle and per trip, premium customer identification index, parking revenue per slot, payment system revenue achieved compared to forecast and addressable market
Intelligent sensor-based infrastructure Data quality index, transportation conditions index, trip time variability indexLow cost efficient, secure and resilient ICT Network load compared to capacity index, network latency, cost of data transfer, network security index
Smart grid, roadway electrification and electric vehicleElectric vehicle charging points per mile, electric vehicle charging points per head of population, number of electric vehicles as a percentage of the total fleet, electric vehicle miles per day, electric vehicle miles per trip, electric vehicle miles between charges
Smart land-use Observed trip generation rates for different land uses, observed actual trips between zones, land value transportation index, zone accessibility index
Strategic business models and partnering Percentage of private sector investment, number of partnerships, improvement in service delivery for each private sector dollar invested
Transportation governanceTransportation efficiency for each dollar spent, supply and demand matching index, transportation agency coordination index, partnership cost-saving index, cost of data storage and manipulation compared to services provided
Transportation managementMobility index, citywide job accessibility index, citywide transportation efficiency index, reliability index, end-to-end time including modal interchanges index
Traveler information Traveler satisfaction index, decision quality information index, behavior change index
Urban analytics Number of analytics in use, value of services managed by analytics, money saved through efficiencies gained by analytics
Urban automationPercentage of automated vehicles within the entire citywide fleet, percentage of automated vehicles in use by city agencies and private fleets, proportion of deliveries made by automated vehicles, proportion of passengers carried by automated transit
Urban delivery and logisticsAverage cost of urban delivery, average time for end-to-end delivery, freight and logistics user satisfaction index, freight management satisfaction index
User focused mobility Citywide mobility index, user satisfaction index, transportation service delivery reliability index
Importance of Operations
• Operations as a significant data generator
• SANDAG• 1 TB per day• Assumed 200 days per year operation• 200 TB per annum
• Connected vehicle• 2 ZB per annum
• The impact of operations on safety, efficiency, and user experience
• Coordination of planning, design, project delivery, operations, and maintenance to deliver quality services
8
Proportion of the data originating
Planning 20%Design 10%Project delivery 5%Operations 50%Maintenance 15%Total 100%
Planning
Design
Project delivery
Operations
Maintenance
Traffic anomaly detection and communications1 Towing and recovery
management2
Mobility as a service10
Transportation Operations Use Case Catalog Version 1
Results driven investment3
Asset management4 Transportation network management5
Transportation systems management and operation impact analysis
6
Developer fee management7 Regionwide safety analysis8 Regionwide speed in
bottleneck analysis9
Connected citizens and travelers11 Project tracking and
coordination12
TSM&O Use Cases
9
Smart Data ManagementNot So Smart Data Management Smart Data Management
Predictive and Prescriptive AnalyticsPetros Xanthopoulos
Analytics
• Descriptive analytics• Prepares and analyzes historical data• Identifies patterns from samples for
reporting of trends• Predictive analytics
• Predicts future probabilities and trends• Finds relationships in data that may not be
readily apparent with descriptive analysis• Prescriptive analytics
• Evaluates and determines new ways to operate
• Targets business objectives• Balances all constraints
”The scientific process of transforming data into insights for the purpose of making better decisions.”
Institute for Operations Research and Management Science (INFORMS)
Descriptive Analytics
• More than just descriptive statistics
Examples
• Foursquare check-ins show the pulse of New York City
https://vimeo.com/75413842
• Data Categories (Labels are not known)
• Used for exploratory/preliminary analysis of the data
Predictive Analytics - Clustering
Stetson Faculty/Staff
UCF faculty/staff
UF undergrad alumni
UF grad colleagues
Classmates from Greece (TUC)
Example Social Media Graph Clustering- LinkedIn
Graphs drawn with SociLab.com
• “Final frontier of analytic capabilities”
• Transform trends and patterns into actionable insights
• Can predictions from Data results in REAL actions???
Prescriptive Analytics
Giving Viewers What They Want
• Fact0: Netflix has 27million subscribers in the US and 33 million worldwide.
• Fact 1: A lot of users had streamed the work of Mr. Fincher, the director of “The Social Network,” from beginning to end.
• Fact 2: films featuring Mr. Spacey had always done well.
• Fact 3: British version of “House of Cards” had done well too.
Giving Viewers What They Want
Potential Benefits in Transportation
McKinsey Global Institute “The age of analytics: competing in a data-driven world” (December 2016)
What Do The Following Have In Common?
• Toilets
• Dog house
• Onesies
• Mattress cover
• Egg tray
They All Can Be Connected To The Internet
The Internet of Things (IoT)
Number Of Devices Connected To The Internet
*from http://www.solidsmack.com/
Evolution of Computer Processors
Evolution of Computer Storage
Big Data
A Data Universe
*from https://www.emc.com/leadership/digital-universe/2014iview/executive-summary.htm
Data Sophistication
Continuous update and time-sensitive queries
become important
OPERATIONALIZINGApplying insights to
transportation operations
Event-based triggering takes hold
ACTIVATINGAutomated transportation
back office
Primarily batch and some ad hoc reports
Increase in ad hoc analysis
ANALYZINGMechanisms
related to transportation demand and
supply
REPORTINGHistorical
Performance Reporting
Analytical modeling
grows
PREDICTINGFuture
Transportation Demand and
Supply
Single View of Transportation – Better, Faster Decisions – Drive Safety, Efficiency, User Experience
Wor
kloa
d C
omp
lexi
ty
Batch
Ad Hoc
ContinuousUpdate/Short Queries
Analytics
Event-basedTriggering
Growth in Query complexity, Workload mixture, Depth of history, Number of users, Expectations
Database Requirement: Analytic foundation must handle multi-dimensional growth!
Machine Learning and TSM&O
28
Summary
• Transportation Systems Management and Operations (TSM&O)
• TSM&O from an analytics perspective
• Predictive and prescriptive analytics
• Machine learning and big data
Discussion